HAL Id: hal-02633630
https://hal.inrae.fr/hal-02633630
Submitted on 27 May 2020
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
against Two Closely Related Pasteurellaceae with
Similar MICs: Differential In Vitro Inoculum Effects
and In Vivo Efficacies.
Guillaume Lhermie, Farid El Garch, Pierre-Louis Toutain, Aude Ferran, Alain
Bousquet-mélou
To cite this version:
Guillaume Lhermie, Farid El Garch, Pierre-Louis Toutain, Aude Ferran, Alain Bousquet-mélou. Bac-terial Species-Specific Activity of a Fluoroquinolone against Two Closely Related Pasteurellaceae with Similar MICs: Differential In Vitro Inoculum Effects and In Vivo Efficacies.. PLoS ONE, Public Library of Science, 2015, 10 (10), pp.e0141441. �10.1371/journal.pone.0141441�. �hal-02633630�
Bacterial Species-Specific Activity of a
Fluoroquinolone against Two Closely Related
Pasteurellaceae with Similar MICs:
Differential
In Vitro Inoculum Effects and In
Vivo Efficacies
Guillaume Lhermie1, Farid El Garch1, Pierre-Louis Toutain2,3, Aude A. Ferran2,3,
Alain Bousquet-Mélou2,3*
1 Vétoquinol, Lure, France, 2 INRA, UMR1331 TOXALIM, Toulouse, France, 3 Université de Toulouse, INPT, ENVT, EIP, UPS, Toulouse, France
*a.bousquet-melou@envt.fr
Abstract
We investigated the antimicrobial activity of a fluoroquinolone against two genetically close bacterial species belonging to the Pasteurellaceae family. Time-kill experiments were used to measure the in vitro activity of marbofloxacin against two strains of Mannheimia haemoly-tica and Pasteurella multocida with similar MICs. We observed that marbofloxacin was equally potent against 105CFU/mL inocula M. haemolytica and P. multocida. However, an inoculum effect was observed with P. multocida, meaning that marbofloxacin activity was decreased against a 108CFU/mL inoculum, whereas no inoculum effect was observed with M. haemolytica. Marbofloxacin activity was also tested in a lung infection model with immu-nocompromised mice intratracheally infected with 109CFU of each bacteria. At the same dose, the clinical and bacteriological outcomes were much better for mice infected with M. haemolytica than for those infected with P. multocida. Moreover, bacteriological eradication was obtained with a lower marbofloxacin dose for mice infected with M. haemolytica. Our results suggest that the differential in vivo marbofloxacin efficacy observed with the two bac-terial species of similar MIC could be explained by a differential inoculum effect. Conse-quently, MICs determined on 105CFU inocula were not predictive of the differences in antibiotic efficacies against high bacterial inocula of closely related bacterial strains. These results could stimulate further investigations on bacterial species-specific antibiotic doses in a clinical setting.
Introduction
In veterinary and human medicine, the fight against antimicrobial resistance requires the reduction of antimicrobials consumption through new regulatory requirements, the
a11111
OPEN ACCESS
Citation: Lhermie G, El Garch F, Toutain P-L, Ferran AA, Bousquet-Mélou A (2015) Bacterial Species-Specific Activity of a Fluoroquinolone against Two Closely RelatedPasteurellaceae with Similar MICs: DifferentialIn Vitro Inoculum Effects and In Vivo Efficacies. PLoS ONE 10(10): e0141441. doi:10.1371/journal.pone.0141441
Editor: Massimiliano Galdiero, Second University of Naples, ITALY
Received: November 28, 2014
Accepted: October 8, 2015
Published: October 27, 2015
Copyright: © 2015 Lhermie et al. This is an open access article distributed under the terms of the
Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability Statement: All relevant data are within the paper.
Funding: The experimental assays were supported by Vetoquinol (Lure, France). The funders (Vetoquinol) had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: Guillaume Lhermie and Farid El Garch have an affiliation to the commercial funder Vetoquinol. This does not alter the authors’
suppression of inappropriate uses, and also through the optimization of dosage regimens for appropriate uses. Such optimization of dosing regimens should more precisely take into account factors depending on the pathogens, the infection sites, the molecules, the hosts, and their interactions [1].
Mannheimia haemolytica and Pasteurella multocida are two Gram negative bacterial species belonging to the Pasteurellaceae family, which are closely genetically related. These bacteria are the most often encountered bacterial aetiological agents of a pulmonary disease in cattle called BRD complex. This disease can be treated by several antibiotics belonging to different families, including fluoroquinolones, and for a given antibiotic the same dosage regimens are registered for the overall disease independently of the bacterial aetiological agent. This situation of “dis-ease claim” is representative of veterinary and human treatments targeting a cluster of patho-gens. In the context of prudent use of antimicrobial drugs, we wondered if such“disease claim” could potentially be improved by more deeply exploring the activity of antibiotics against the different aetiological agents.
The efficacy of an antibiotic treatment depends among diverse factors on the susceptibility of the targeted pathogens, and it is the reason why the MICs of antibiotics used against bacteria are combined with pharmacokinetic parameters to explore the efficacy of dosage regimens. For fluoroquinolones, two PK/PD indices are correlated with the effectiveness of an antimicrobial treatment: AUC/MIC and Cmax/MIC [2,3]. However, a set of studies have pointed out the limitations of MIC as the sole indicator to capture the complexity of interactions between bugs and drugs [4]. These limitations are linked to the standard procedure for its determination that consists in assessing antibiotic activity on an inoculum of 5.105CFU/mL in the presence of constant concentrations of drugs and after a fixed time of incubation [5]. As an example, the standard MIC determination does not take into account the inoculum effect whereas several studies have stressed that the activity of antibiotics is decreased when they face high bacterial inocula at the infection sites [6–10].
By using in vitro and in vivo assays, the aim of this study was to determine the activity of marbofloxacin, a fluoroquinolone used in veterinary medicine in Europe, against two strains of two closely related bacterial species belonging to the same Familiae (Pasteurellaceae) that have similar MICs and are targeted in the field by the same dosing regimens.
Material and Methods
Bacterial strains, media, antibiotic and reagents
A strain of M. haemolytica serotype A2, isolated in 2010 in the Netherlands from the lungs of a calf with clinical signs of BRD, and a strain of P. multocida isolated in a diseased pig in France were used in this study.
Mueller Hinton (MH) broth and agar and Brain Heart Infusion Broth (BHIB) were pur-chased from Biomerieux.
Marbofloxacin was provided by Vetoquinol SA and used for MIC and MBC determinations and for time-kill experiments. For in vitro testing, the powder was diluted in sterile distilled water from a stock solution at 1 mg/mL stored at -20°C. For in vivo testing (mice challenge), a commercial formulation of marbofloxacin for veterinary use approved in France (Marbocyl 10%1) was used immediately after dilution in sterile distilled water.
In vitro testing
Determination of the Minimum Inhibitory Concentration (MIC) and the Minimum Bactericidal Concentration (MBC). MIC were determined in triplicate according to the CLSI reference methods (5). MIC were also determined in Brain Heart Infusion Broth (BHIB)
adherence to PLOS ONE policies on sharing data and materials.
with bacterial inocula of 5.105CFU/mL, using the same microdilution method since BHIB gave better growth of both strains and was used for killing curves experiments.
Bactericidal effects were determined with bacterial inocula of 105CFU/mL and 108CFU/ mL, from time kill experiments by subculturing in MH agar plates, where wells with no visible growth were observed. The minimum bactericidal concentration (MBC) was defined as the lowest marbofloxacin concentration giving a decrease of more than 3log10 CFU/mL (99.9%) at 24h.
Time kill experiments. After overnight growth on MH agar plates, inocula of M. haemoly-tica and P. multocida were prepared by transferring several colonies into fresh BHIB to obtain suspensions containing 105CFU/mL and 108CFU/mL. Suspensions were incubated for one hour at 37°C to reach exponential growth phase before adding marbofloxacin at different con-centrations (0, 0.25, 0.5, 0.75, 1, 2, 4, 8, 16 and 32 fold the MIC). Aliquots of 500μL were removed from bacterial suspensions at 0, 0.5, 1, 2, 4, 6, 8 and 24h after incubation at 37°C. They were centrifuged at 3000g for 10 minutes and pellets were then resuspended in 500μL of 0.9% NaCl to eliminate the carryover of marbofloxacin. TenμL of successive 10-fold dilutions were plated on MH agar. The colonies were counted after 24h of incubation at 37°C. The detec-tion limit of colony counting was 100 CFU/mL.
Mean Inoculum growth: From each time-kill curve, we calculated the Area Under the Curve (AUC) of bacterial counts from time 0 to 24h. This AUC was divided by 24 hours to obtain an average inoculum size over the investigated time interval (24 hours) for each x tested marbo-floxacin concentration [3]. The basal inoculum (IBASAL) was equal to the inoculum size before
exposure to marbofloxacin. Mean inoculum growth (ΔI(x)) was then defined byEq 1:
DIðxÞ¼ IxIBASAL ð1Þ
whereΔI(x) is expressed in log CFU/mL, Ix(log CFU/mL) is the mean inoculum obtained after
exposure to a given marbofloxacin concentration x and IBASAL(log CFU/mL) is the inoculum
size before exposure to marbofloxacin.
Modelling of mean inoculum growth: Mean inoculum growth was modelled as a function of marbofloxacin concentration, using a sigmoid dose-response effect model implemented in the WinNonlin software version 5.3 (Pharsight, Moutain View). We re-parameterized equations previously used elsewhere [11,12] which are derived from a classical sigmoid Emaxmodel, to
obtain Eqs2and3allowing to directly calculate two parameters of interest, i.e. EC-3logthe
anti-microbial concentration producing a reduction of 3log10(99.9% reduction) of the basal
inocu-lum of 105CFU/mL, and EC-5logthe antimicrobial concentration producing a reduction of
5log10(99.999% reduction) of the basal inoculum of 108CFU/mL.
DIðxÞ¼DIMAX
ðDIMAX DIMINÞ x=EC3log
g ð3 DIMIN
=
Þ ðDIMAXþ3Þ þ x= EC3log g ð2Þ DIðxÞ ¼ DIMAXðDIMAX DIMINÞ x=EC5log
g ð5 DIMIN
=
Þ ðDIMAXþ5Þ þ x= EC5log g ð3ÞwhereΔI(x) (log CFU/mL) is the mean inoculum growth for marbofloxacin concentration x, ΔIMAXandΔIMIN(log CFU/mL) the maximal and minimal inoculum growth, obtained without
(μg/mL) the tested marbofloxacin concentration, EC-3log(μg/mL) the antimicrobial
concentra-tion producing a reducconcentra-tion in the basal inoculum of 3 log10, i.e.ΔI(EC-3log) = -3, EC-5log(μg/
mL) the antimicrobial concentration producing a reduction in the basal inoculum of 5 log10, i.
e.ΔI(EC-5log) = -5, andγ the sigmoid coefficient of the curve. These values of the mean
inocu-lum changes (-3 log10and -5 log10) were chosen to take into account the size of the initial
inoc-ulum (5 log10and 8 log10). In both cases, the remaining inoculum (2 or 3 log10) can be
considered as similar, and low enough to be eradicated in vivo by the host natural defences.
In vivo testing
Animals. All animal procedures were carried out in accordance with international accepted standards of animal care under agreement number TOXCOM/0022/AF GL (Ethic committee: Comité d’éthique de Pharmacologie‐Toxicologie de Toulouse‐Midi Pyrénées, France #C2EA-86) for animal experimentation of French Ministry of Agriculture.
Eighty female Swiss mice weighing 20 to 30g (Charles River Laboratories, L’Arbresle, France) and aged from 8 to 12 weeks were conditioned in our laboratory at least for one week before experimentation. Animals were housed with a 12 hours day-night cycle in a temperature (22°C) controlled room in autoclavable type II L cages with filters on Souralit1(hygiene ani-mal bedding) and had water and food (Tecklad global diet 2014 Harlan1) access ad libitum.
Induction of neutropenia. Cyclophosphamide (Sigma-Aldrich, St Quentin Fallavier, France) was administered intraperitoneally to mice at doses of 150 mg/kg 4 days before inocu-lation, and 100 mg/kg one day before inoculation.
Lung infection model. Mice were purchased specific pathogen free and the health status was checked twice a day before and during experiments. We verified mobility, respiratory movements and piloerection. If immobility was observed, we euthanized animals with the humane method described below. These animals were included in the“dead group”. If piloer-ection and/or increase of respiratory movements were detected, the mice were more carefully observed during the day.
On day 0, mice were anesthetized by an intraperitoneal injection of a mixture of 0.17 mg/kg medetomidine (Domitor1, Zoetis, France) and 80 mg/kg ketamine (Chlorketam1, Vétoqui-nol, France).
The tracheas were cannulated with a 22 G catheter without mandrel. The lungs were inocu-lated with 20μL of M. haemolytica or P. multocida suspension containing 5x1011CFU/mL,
cor-responding to a total of 1x109CFU/lung. To constitute a control, 2 mice remained uninfected and were isolated in a separate box.
Three hours after bacterial challenge, mice received a single intraperitoneal injection of mar-bofloxacin. Eight groups of 10 to 13 mice were randomly constituted and were treated with 0, 1, 5, 10, 20 or 40 mg/kg of marbofloxacin for the mice inoculated with M. haemolytica, and with 0 or 40 mg/kg of marbofloxacin for the mice inoculated with P. multocida. For ethical rea-sons, only 4 to 8 mice were allocated in groups with low variability.
Clinical signs of disease and mortality were recorded over 48h, and at day 2 after challenge, all surviving mice were sacrificed humanely by an intraperitoneal injection of 120 mg/kg pen-tobarbital sodium (Dolethal1, Vétoquinol, France). The lungs were aseptically removed and homogenized in 5mL of 0.9% NaCl. The homogenates were centrifuged at 3000 g for 10 min-utes and the pellets resuspended in 2 mL of 0.9% NaCl. TenμL of successive 10-fold dilutions were then plated on MH agar. The bacterial counts per lung were recorded after 24h of incuba-tion at 37°C. The detecincuba-tion limit of colony counts was 200 CFU/lung. Mice were then divided into 3 categories regarding the effect of marbofloxacin:“persistent infection” (decrease in bac-terial load<3log10 CFU in lungs), “bactericidal effect” (decrease in bacterial load >3log10
CFU in lungs),“bacterial eradication” (bacterial counts below the LOQ). In the follow-up, dead mice were classified into the“persistent infection” category.
Statistical methods. The bacterial counts in lungs after 48 h were compared with an analy-sis of variance followed by a Dunnett comparison.
Results
In vitro testing
MIC and MBC results. For both bacterial species the values of MIC, MBC and the MBC/ MIC ratio are shown inTable 1. The MIC values of marbofloxacin were 0.03μg/mL for M. hae-molytica, and 0.016μg/mL for P. multocida. These values were identical in MHB and BHIB.
The MBC values observed with the standard inoculum of 5.105CFU/mL showed very slight differences, i.e. equal or less than one 2-fold dilution, compared to the corresponding MIC val-ues for both M. haemolytica (0.023μg/mL) and P. multocida (0.03 μg/mL), as indicated by MBC/MIC ratios (Low inoculum,Table 1). The MBC value obtained with the 5.108CFU/mL inoculum was equal to the MIC (0.03μg/mL) for M. haemolytica, but was 8-fold the MIC (0.12μg/mL) for P. multocida (High inoculum,Table 1).
Time-kill curves. The time-kill curves of M. haemolytica and P. multocida exposed to dif-ferent concentrations of marbofloxacin are presented inFig 1.
For M. haemolytica, we obtained a concentration-dependant bactericidal effect. For the low inoculum of 105CFU/mL the MBC was equal to 0.75xMIC. Marbofloxacin killed 99.9% of the bacteria within 0.5h at concentrations higher than 2xMIC, within 2h at 1 xMIC and within 4h at 0.75xMIC.
For the high inoculum (108CFU/mL), marbofloxacin killed 99.9% of the bacteria within 0.5h at concentrations 4xMIC, within 1h at 2xMIC, and within 2h at 1xMIC.
For P. multocida, the results obtained with the low inoculum of 105CFU/mL were similar to those observed with M. haemolytica. The MBC of marbofloxacin was equal to 1xMIC, even if the effect was slower than for M. haemolytica (killing<3log10 at 8h). In addition, marbofloxa-cin achieved 99.9% killing within 0.5, 1 and 2h at concentrations of 8x, 4x and 2xMIC, respec-tively. For the high inoculum of 108CFU/mL P. multocida, the MBC was 8xMIC;
marbofloxacin achieved 99.9% killing within 1h at concentrations 16xMIC, 2h at 8xMIC. At 4xMIC, a transitory bactericidal activity was observed for the first 8 hours but a regrowth was observed at 24h.
Table 1. Marbofloxacin MIC and MBC determined in BHIB for M. haemolytica and P. multocida.
MIC (μg/mL) MBC (μg/mL) MBC/MIC ratio
M. haemolytica 0.03 low inoculum 0.023 0.75 high inoculum 0.03 1 P. multocida 0.016 low inoculum 0.03 2 high inoculum 0.12 8
MIC were determined with a standard inoculum of 5.105CFU/mL. MBC were extracted from time-kill experiments.
Low inoculum: 5.105CFU/mL. High inoculum: 5.108CFU/mL.
Pharmacodynamic modelling. The relations between inoculum changes and marbofloxa-cin concentrations are presented inFig 2. The observed data were fitted with the models described by Eqs2and3and the pharmacodynamic parameters are presented inTable 2.
For the low inoculum of both M. haemolytica and P. multocida, concentration-effect curves were close, with marbofloxacin concentrations associated withΔI = -3 around 1xMIC (seeFig 2, upper panel). Indeed, data fitting usingEq 2indicated that EC-3log(concentrations associated
withΔI = -3) were 0.69xMIC and 1.34xMIC for M. haemolytica and P. multocida, respectively. The ratio of EC-3logfor P. multocida and M. haemolytica was 1.95 (Table 2).
For the high inoculum, the concentration-effect curve of M. haemolytica (Fig 2, lower panel) showed that inoculum changes occurred in a concentration range of marbofloxacin sim-ilar to that observed with the low inoculum of M. haemolytica (between 0.5 and 2- fold the MIC). Conversely, the concentration-effect curve for the high inoculum of P. multocida was dramatically shifted to the right, indicating that the marbofloxacin concentrations required to achieve the same magnitude of inoculum change were much higher with P. multocida than with M. haemolytica. Data fitting usingEq 3indicated that EC-5log(concentrations associated
Fig 1. Time-kill curves of low (105CFU/mL) (A and B) and high (108
CFU/mL) (C and D) inocula of M. haemolytica (A and C) and for P. multocida (B and D) in the presence of marbofloxacin (concentrations shown in multiple of the MIC (MIC = 0.016μg/mL for P multocida and MIC = 0.03 μg/mL for M. haemolytica).
withΔI = -5) were 1.03xMIC and 10.0xMIC for M. haemolytica and P. multocida, respectively (Table 2). The ratio of EC-5logfor P. multocida and M. haemolytica was 9.74 (Table 2).
Infection model
Clinical outcomes: mouse survival. Mice that were not infected showed normal activity and no death, during the whole experiment. The results obtained in the different groups of infected mice are presented inTable 3.
Fig 2. Bactericidal effect of marbofloxacin against low (A) and high (B) inocula of M. haemolytica (full symbols) and P. multocida (open symbols) isolates.
In the two infected/non-treated groups (0 mg/kg marbofloxacin dose), all the mice infected with M. haemolytica or P. multocida presented clinical signs of disease. About 15% (2/13) of the mice infected with M. haemolytica died between 36–48 h post-challenge, and 100% (4/4) of the mice infected with P. multocida died between 12–36 h post-challenge (85% and 0% survival at 48 h, respectively,Table 3).
In the group of mice infected with P. multocida and treated with the 40 mg/kg marbofloxa-cin dose, 100% (11/11) of the mice showed persistent clinical signs of infection (they developed mildly scruffy coat and were prostrate) and were still alive at 48h post-challenge (Table 3).
In the groups of mice infected with M. haemolytica and treated with the 40 mg/kg marbo-floxacin dose, 100% survival was observed at 48 h post-challenge and none of the mice showed any clinical signs of disease. When using decreasing doses of marbofloxacin (20, 10, 5 and 1 mg/kg) in mice infected with M. haemolytica, 100% survival was observed at 48 h post-chal-lenge and the percentage of mice presenting clinical signs progressively increased from 20% to 100% (Table 3).
Microbiological outcomes: pulmonary bacterial loads. The lung bacterial counts mea-sured in mice 48h after the challenge are reported inFig 3.
Table 2. Pharmacodynamic parametersa
describing the antimicrobial activity of marbofloxacin against different inoculum sizes of M. haemolytica (MH) and P. multocida (PM) in BHIB.
Basal inoculum Bacteria EC-3logbor EC-5logc PM/MH γd ΔIMAXe ΔIMINf
(Log10CFU/mL) (MIC multiple) ratio (Log10CFU/mL) (Log10CFU/mL)
5 MH 0.69±0.05b 9.14±2.12 3.15±0.12 -3.50±0.09
PM 1.34±0.29b 1.95 2.53±1.06 2.82±0.69 -4.49±0.60
8 MH 1.03±0.02c 4.17±0.21 0.67±0.10 -6.28±0.04
PM 10.0±1.53c 9.71 1.34±0.32 0.57±0.35 -6.70±0.65
aMean±SE of parameters calculated with Eqs2and3(sigmoid E
maxmodel) bAntibiotic concentration reducing by 3 log
10the basal inoculum of 105CFU/mL cAntibiotic concentration reducing by 5 log
10the basal inoculum of 108CFU/mL dSigmoid coefficient of the curve.
eMaximal growth compared to the basal inoculum.
fMinimal growth (maximal killing) compared to the basal inoculum.
doi:10.1371/journal.pone.0141441.t002
Table 3. Effect of the different marbofloxacin doses on the survival and clinical and bacteriological cure of mice infected with P. multocida or M. haemolytica.
M. haemolytica P. multocida
Marbofloxacin dose (mg/kg) 0 1 5 10 20 40 0 40
Number of mice 13 4 10 12 12 8 4 11
% of mice with clinical signs 100 100 60 58 20 0 100 100
% of surviving mice 85 100 100 100 100 100 0 100
% of mice with bacterial eradication 0 0 10 25 33 37 0 0
% of mice with bactericidal effect 0 0 50 33 58 63 0 9
% of mice with persistent infection 100 100 40 42 9 0 0 9
Bacterial eradication: bacterial counts below the LOQ; bactericidal effect: decrease in bacterial load>3log10 CFU; persistent infection: decrease in bacterial load<3log10 CFU.
The mean bacterial counts in the lungs of untreated animals infected with M. haemolytica and P. multocida were 8.48±0.97 and 10.08±0.06 log10CFU/lung, respectively.
The 40 mg/kg marbofloxacin dose led to a dramatic decrease in the mean bacterial counts for M. haemolytica infected mice (1.91±0.90 log10CFU/lung) compared to the untreated group
(8.48±0.97 log10CFU/lung), whereas the mean bacterial counts were only slightly decreased
for P. multocida infected mice (8.10±1.30 versus 10.08±0.06 log10CFU/lung). For M.
haemoly-tica infected mice, decreasing the marbofloxacin doses from 20 to 1 mg/kg showed a progres-sive increase in the mean bacterial loads in the lungs at 48h post-challenge. An analysis of variance followed by a Dunnett comparison indicated that the mean bacterial count for 1 mg/ kg was not significantly different (p>0.05) from the untreated group, whereas the mean bacte-rial counts were significantly decreased compared to the control group for the 5 mg/kg and higher doses (p<0.05).
The percentages of mice exhibiting“persistent infection” (decrease in bacterial load <3log10 CFU), “bactericidal effect” (decrease in bacterial load >3log10 CFU), or “bacterial eradication” (bacterial counts below the LOQ) are presented inTable 3.
The 40 mg/kg marbofloxacin dose produced bacterial eradication or a bactericidal effect in 37% and 63% of mice infected with M. haemolytica, respectively. In mice infected with P. mul-tocida, the 40 mg/kg marbofloxacin dose produced no bacterial eradication, and only a 9% bac-tericidal effect.
In mice infected with M. haemolytica, the decrease in marbofloxacin doses from 20 to 1 mg/ kg was correlated with an increase in the percentage of mice with persistent infection: 9%, 42%, 40% and 100% for the marbofloxacin doses of 20, 10, 5 and 1 mg/kg, respectively.
Discussion
In the present study, we assessed in a mouse model of pulmonary infection the efficacy of an antimicrobial treatment with marbofloxacin against two bacterial species belonging to the Pas-teurellaceae family: M. haemolytica and P. multocida. These two bacterial species are among the etiologic agents of the same disease in cattle.
Our main result demonstrated that the dose of marbofloxacin required to cure experimental pneumonia in mice was dramatically different depending on the bacterial pathogen responsible for the infection, despite the fact that these pathogens have similar MICs and are closely
Fig 3. Mean bacterial counts in the lungs of mice infected with P. multocida (dark bars) or with M. haemolytica (white bars) after intraperitoneal administrations of different doses of marbofloxacin. Different letters in superscript above bars indicate that values are statistically different.
genetically related (Pasteurellaceae family). Moreover, in vitro assessment of the antibacterial activity of marbofloxacin against both bacterial species demonstrated that the inoculum effect, that corresponds to the decrease of antibacterial activity when inoculum size increases, and which was previously identified as an important factor of antimicrobial efficacy of fluoroquino-lone [9,13,14], was also dramatically different in magnitude between the two bacterial species.
Our first objective was to develop a rodent model of pulmonary infection with M. haemoly-tica or P. multocida to enable clinical and bacteriological outcomes to be assessed after marbo-floxacin treatments. In our experiments, we chose to depress the immune response of the mice, by the administration of cyclophosphamide before the course of challenges [15] to get similar bacterial counts at the time of treatment with both bacteria. By minimizing the implication of the natural (specific and non-specific) defence, neutropenia enabled overt disease to be estab-lished in order to compare the effects of marbofloxacin on both bacterial species.
In this pneumonia model, we first evaluated the clinical and microbiological outcomes after marbofloxacin treatment at a dose of 40 mg/kg. This dose had already been tested in a pneu-monia model with P. multocida in immunocompetent mice [16]. In the present study with immunocompromised mice infected by P. multocida, the 40 mg/kg dose prevented mortality at 48 h post-challenge, but it produced neither a clinical cure nor a significant decrease in the bac-terial loads in the lungs. Conversely, the 40 mg/kg dose in immunocompromised mice chal-lenged with M. haemolytica was associated with a 100% clinical cure rate and a significant microbiological effect in all treated mice (including a microbiological cure rate of 40%). We then decided to explore the magnitude of this differential drug efficacy by treating M. haemoly-tica infected mice with decreasing doses of marbofloxacin. For ethical reasons, as the 40 mg/kg dose was poorly effective against P. multocida, we did not test lower doses of marbofloxacin in P. multocida infected mice. After testing marbofloxacin doses of 20, 10, 5 and 1 mg/kg, we demonstrated that the 5 mg/kg dose was the lowest dose producing better clinical and microbi-ological outcomes in the M. haemolytica pneumonia than the 40 mg/kg marbofloxacin dose in the P. multocida pneumonia (Fig 3andTable 3).
We explored in vitro using time-kill experiments the antibacterial activity of marbofloxacin against both P. multocida and M. haemolytica strains. The bacteriological results were
expressed in terms of mean inoculum changes over 24 h, and analysed as a function of marbo-floxacin concentrations [11,12]. With initial inocula of 105CFU/mL, the marbofloxacin con-centrations producing a 99.9% (-3 log10) reduction in the basal inoculum were similar between
the two bacterial strains and close to their respective MICs (0.69xMIC for M. haemolytica and 1.34xMIC for P. multocida,Table 2). Similar results were obtained by Ferran et al. [12] with P. multocida (2.06xMIC) and E. coli (1.80xMIC). This is in agreement with the observation that the MBC of marbofloxacin was close to the MIC for both P.multocida and M. haemolytica. Altogether, these results support the view that MIC is a relevant indicator of the marbofloxacin bactericidal effect against low bacterial inocula.
Using initial inocula of 108CFU/mL, huge differences were observed when comparing the marbofloxacin activity against P. multocida and M. haemolytica strains (Fig 2, lower panel;
Table 2).
For P. multocida, the concentration-response curve was shifted to the right of the x-axis, and the marbofloxacin concentration that gave a 99.999% (-5 log10) reduction in the basal
inoc-ulum was about 10-fold the MIC. These results are evidence of the inocinoc-ulum effect, consisting of a reduction in the antibiotic activity while bacterial inoculum size increases. This effect was previously demonstrated with P. multocida [12] and with other Gram-negative and Gram-pos-itive bacterial species [9,13,14,17].
Unexpectedly, the concentration-response curve for M. haemolytica remained close to that obtained with the 105CFU/mL inoculum, and the marbofloxacin concentrations associated
with a 99.999% (-5 log10) reduction in the basal inoculum was 1.03xMIC which is similar to
the value obtained with a -3 log10reduction of the 105CFU/mL inoculum (0.69xMIC). The
same result was obtained with another strain of M. haemolytica (data not shown). M. haemoly-tica is, to our knowledge, the first reported bacterial species that does not exhibit such an inoc-ulum effect for a fluoroquinolone.
Interestingly, the results obtained in vitro with the high inocula (108CFU/mL) of P. multo-cida and M. haemolytica (Fig 2) mirror those obtained in vivo in the pneumonia model pro-duced with a bacterial load of 109CFU/lung. Indeed, the ratio of in vitro effective
concentrations (EC-5log) for P. multocida and M. haemolytica was equal to 9.71 (Table 2), and
the ratio of in vivo doses producing similar responses for P. multocida and M. haemolytica was about 8 (40 mg/kg for P. multocida and 5 mg/kg M. haemolytica).
In a set of published works, it has been demonstrated in experimental models of infections that the antimicrobial activity of fluoroquinolones or third-generation cephalosporins could be reduced against high bacterial loads compared to low bacterial loads of different bacterial spe-cies, including Escherichia coli, Pasteurella multocida, Klebsiella pneumoniae and Staphylococ-cus aureus [6,9,13,18]. In all these cases, in vitro time-kill experiments have shown the occurrence of an inoculum effect.
Altogether, these data suggest that the higher activity of marbofloxacin in the pneumonia induced by a high inoculum of M. haemolytica compared to P. multocida was potentially linked to the absence of an inoculum effect in M. haemolytica. Therefore, it might be speculated that marbofloxacin was as effective against the high inoculum of M. haemolytica as it would be against a low inoculum. Unfortunately, this situation could not be tested in our model of infec-tion because of the low pathogenicity of M. haemolytica in mice, which did not produce overt clinical symptoms of disease with a low bacterial load.
In conclusion, our study is the first to our knowledge demonstrating that the optimal antibi-otic dose of a fluoroquinolone was markedly different between two bacterial species that exhibit similar MIC and are closely genetically related. The appropriate and prudent use of antimicro-bials is a growing challenge for both human and veterinary medicine, and as already suggested, the optimization of antibiotic dosing regimens should take all its place in this challenge. In this context, our present results could stimulate further research to investigate the potential imple-mentation of bacterial species-specific antibiotic doses in a clinical setting.
Author Contributions
Conceived and designed the experiments: GL FEG PLT AAF ABM. Performed the experi-ments: GL FEG AAF. Analyzed the data: GL FEG PLT AAF ABM. Contributed reagents/mate-rials/analysis tools: GL FEG PLT AAF ABM. Wrote the paper: GL FEG PLT AAF ABM.
References
1. Canton R, Morosini MI. Emergence and spread of antibiotic resistance following exposure to antibiotics. FEMS Microbiol Rev. 2011; 35: 977–991. PMID:21722146. doi:10.1111/j.1574-6976.2011.00295.x
2. Drusano GL, Preston SL, Fowler C, Corrado M, Weisinger B, Kahn J. Relationship between fluoroquin-olone area under the curve: minimum inhibitory concentration ratio and the probability of eradication of the infecting pathogen, in patients with nosocomial pneumonia. J Infect Dis. 2004; 189: 1590–1597. PMID:15116294.
3. Forrest A, Nix DE, Ballow CH, Goss TF, Birmingham MC, Schentag JJ. Pharmacodynamics of intrave-nous ciprofloxacin in seriously ill patients. Antimicrob Agents Chemother. 1993; 37: 1073–1081. PMID:
8517694.
4. Gehring R, Riviere JE. Limitations of MIC as the sole criterion in antimicrobial drug dosage regimen design: the need for full characterization of antimicrobial pharmacodynamic profile especially for drug-resistant organisms. Vet J. 2013; 198: 15–18. PMID:24025153. doi:10.1016/j.tvjl.2013.07.034
5. CLSI. Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria That Grow Aerobically; Approved Standard-Eighth Edition. Document M7-A8. 2009; 29-2(ISBN 1-56238-689-1).
6. Ferran AA, Kesteman AS, Toutain PL, Bousquet-Melou A. Pharmacokinetic/pharmacodynamic analy-sis of the influence of inoculum size on the selection of reanaly-sistance in Escherichia coli by a quinolone in a mouse thigh bacterial infection model. Antimicrob Agents Chemother. 2009; 53: 3384–3390. PMID:
19487439. doi:10.1128/AAC.01347-08
7. Jumbe N, Louie A, Leary R, Liu W, Deziel MR, Tam VH, et al. Application of a mathematical model to prevent in vivo amplification of antibiotic-resistant bacterial populations during therapy. J Clin Invest. 2003; 112: 275–285. PMID:12865415.
8. Kesteman AS, Ferran AA, Perrin-Guyomard A, Laurentie M, Sanders P, Toutain PL, et al. Influence of inoculum size and marbofloxacin plasma exposure on the amplification of resistant subpopulations of Klebsiella pneumoniae in a rat lung infection model. Antimicrob Agents Chemother. 2009; 53: 4740– 4748. PMID:19738020. doi:10.1128/AAC.00608-09
9. Mizunaga S, Kamiyama T, Fukuda Y, Takahata M, Mitsuyama J. Influence of inoculum size of Staphy-lococcus aureus and Pseudomonas aeruginosa on in vitro activities and in vivo efficacy of fluoroquino-lones and carbapenems. J Antimicrob Chemother. 2005; 56: 91–96. PMID:15890721.
10. Vasseur MV, Laurentie M, Rolland JG, Perrin-Guyomard A, Henri J, Ferran AA, et al. Low or high doses of cefquinome targeting low or high bacterial inocula cure Klebsiella pneumoniae lung infections but differentially impact the levels of antibiotic resistance in fecal flora. Antimicrob Agents Chemother. 2014; 58: 1744–1748. PMID:24395228. doi:10.1128/AAC.02135-13
11. Regoes RR, Wiuff C, Zappala RM, Garner KN, Baquero F, Levin BR. Pharmacodynamic functions: a multiparameter approach to the design of antibiotic treatment regimens. Antimicrob Agents Chemother. 2004; 48: 3670–3676. PMID:15388418.
12. Ferran AA, Bibbal D, Pellet T, Laurentie M, Gicquel-Bruneau M, Sanders P, et al. Pharmacokinetic/ pharmacodynamic assessment of the effects of parenteral administration of a fluoroquinolone on the intestinal microbiota: comparison of bactericidal activity at the gut versus the systemic level in a pig model. Int J Antimicrob Agents. 2013; 42: 429–435. PMID:24021905. doi:10.1016/j.ijantimicag.2013. 07.008
13. Ferran A, Dupouy V, Toutain PL, Bousquet-Melou A. Influence of inoculum size on the selection of resistant mutants of Escherichia coli in relation to mutant prevention concentrations of marbofloxacin. Antimicrob Agents Chemother. 2007; 51: 4163–4166. PMID:17709467.
14. Udekwu KI, Parrish N, Ankomah P, Baquero F, Levin BR. Functional relationship between bacterial cell density and the efficacy of antibiotics. J Antimicrob Chemother. 2009; 63: 745–757. PMID:19218572. doi:10.1093/jac/dkn554
15. Andes D, Craig WA. Pharmacodynamics of the new fluoroquinolone gatifloxacin in murine thigh and lung infection models. Antimicrob Agents Chemother. 2002; 46: 1665–1670. PMID:12019073. 16. Ferran AA, Toutain PL, Bousquet-Melou A. Impact of early versus later fluoroquinolone treatment on
the clinical; microbiological and resistance outcomes in a mouse-lung model of Pasteurella multocida infection. Vet Microbiol. 2011; 148: 292–297. PMID:20888712. doi:10.1016/j.vetmic.2010.09.005
17. Morrissey I, George JT. The effect of the inoculum size on bactericidal activity. J Antimicrob Chemother. 1999; 43: 423–425. PMID:10223603.
18. Kesteman AS, Perrin-Guyomard A, Laurentie M, Sanders P, Toutain PL, Bousquet-Melou A. Emer-gence of resistant Klebsiella pneumoniae in the intestinal tract during successful treatment of Klebsiella pneumoniae lung infection in rats. Antimicrob Agents Chemother. 2010; 54: 2960–2964. PMID:
